Human autoimmune diseases have a striking genetic association with particular alleles of major histocompatability complex (“MHC”) class I or class II genes. The field was established by the seminal discovery of HLA-B27 linked susceptibility to ankylosing spondylitis, a chronic inflammatory joint disease (Brewerton (et al., 1973; Schlosstein et al., 1973). MHC associated susceptibility has now been documented for a variety of human autoimmune diseases, including insulin dependent diabetes mellitus (IDDM), rheumatoid arthritis (RA), pemphigus vulgaris (PV), multiple sclerosis (MS) and myasthenia gravis (MG), just to name a few (Todd et al., 1987; Ahmed et al., 1990; Ahmed et al. 1991; Lanchbury & Panayi, 1991; Spielman & Nathenson, 1982; Protti et al., 1993).
The MHC locus most commonly associated with autoimmune disease is the HLA-DRB locus (also known as DRB1), a highly polymorphic locus with over fifty known alleles. For example, a large body of epidemiological work has documented the association of rheumatoid arthritis with the DR4 (DRB1*0401, DRB1*0404) and DR1 (DRB1*0101) alleles, with the DR4 alleles conferring a higher risk than DR1 (Lanchbury & Panayi, 1991). The risk is dramatically increased when the subject is homozygous or heterozygous for DRB1*0401 and/or DRB1*0404. The observation that arthritis is associated with three DR alleles that are structurally similar led to the development of the ‘shared epitope’ hypothesis as DRB1*0401, 0404 and 0101 share critical polymorphic residues in the DRβ67–71 cluster (Gregersen et al. 1987; Lanchbury & Panayi, 1991). These residues (in particular DRβ71) appear to be critical in defining the selectivity of peptide binding to the disease associated molecules.
Pemphigus vulgaris (PV) is an autoimmune disease of the skin in which high titer auto-antibody production to an epidermal cell adhesion molecule (desmoglein 3) results in a loss of keratinocyte adhesion (acantholysis) and severe blister formation (Amagai et al., 1991). In different ethnic groups the disease is associated either with a DR4 allele (DRB1*0402) or with a rare DQ1 allele (DQB1*05032); only a small fraction of PV patients have neither susceptibility gene (Ahmed et al., 1991; Ahmed et al., 1990; Scharf et al., 1988). The DR4 subtype associated with pemphigus differs only at three positions in the DRβ67–71 cluster from the DR4 subtype associated with RA. The PV associated molecule has a negative charge (Glu) at the critical position (DRβ71); the neighboring position (DRβ70) is also negatively charged. The DR4 subtype associated with PV is the only one that carries a negative charge at DRβ71; a positive charge (Arg) is found at DRβ71 in the RA associated DR4 molecules.
Efforts to identify sequence homologies between self-peptide epitopes that might be involved in autoimmunity and various bacterial and viral pathogens have therefore been made. These homology searches have focused on alignments with sequence identity. No success has been reported using such alignments in identifying epitopes from pathogens that could cross react with presumably pathogenic T cell lines from human patients with autoimmune disease (Oldstone, 1990). A sequence identity was recently found between an epitope in a Coxsackie virus protein and GAD65, suspected of being an autoantigen in diabetes. These peptides could reciprocally generate polyclonal T cell lines from mice that cross react with the other peptides (Tian, et al., 1994). No evidence, however, was provided that these peptides could stimulate clones from diabetic mice (or humans).
Recent developments in the field, in particular the identification of allele specific peptide binding motifs have transformed the field (Madden et al., 1991; Rotschke & Falk, 1991). Based on this knowledge the structural basis for MHC linked susceptibility to autoimmune diseases can be reassessed at a level of detail sufficient for solving longstanding questions in the field. Motifs for peptide binding to several MHC class I and class II molecules have been defined by sequence analysis of naturally processed peptides and by mutational analysis of known epitopes. MHC class I bound peptides were found to be short (generally 8–10 amino acids long) and to possess two dominant MHC anchor residues; MHC class II bound peptides were found to be longer and more heterogeneous in size (Madden et al., 1991; Rotschke & Falk, 1991; Jardetzky et al. 1991, Chicz et al. 1993). Due to the size heterogeneity, however, it has proven more difficult to define MHC class II binding motifs based on sequence alignments. More recently, a crystal structure for HLA-DR1 demonstrated that there is a dominant hydrophobic anchor residue close to the N-terminus of the peptide and that secondary anchor residues are found at several other peptide positions (Brown et al., 1993). Even this work, however, could not provide a detailed description of the binding pockets of HLA-DR proteins, the particular residues involved in the formation of these pockets of the structural requirements or antigens for MHC binding.
In the present disclosure, a detailed description of the HLA-DR antigen binding pockets is provided (Stem et al., 1994). With this information, together with functional information defining those amino acids of the self or non-self antigen that are needed for MHC binding and TCR contact (e.g., Wucherpfennig et al. 1994a,), binding motifs for the various HLA-DR allotypes may be developed, self epitopes involved in autoimmune disease may be identified and a method is provided for identifying bacterial and viral epitopes which may initiate a human autoimmune response.